To elucidate the characteristics of air density over the Qinghai-Tibet Plateau and their impacts on wind turbine performance, a field experiment was conducted at a plateau wind farm in Dulan County, Qinghai Province, employing a nacelle-mounted LiDAR system and a 3.3 MW horizontal axis wind turbine. Based on in situ measurements of temperature, barometric pressure, relative humidity, and wind speed, the air density was calculated, and its temporal variation and probability distribution were analyzed. Furthermore, the effects of different density classes on power output and rotor speed were investigated. The results show that air density exhibits significant diurnal periodic fluctuations, with a mean value of only 0.902 kg·m-3, representing a 26.37% reduction relative to the standard atmospheric density. The probability distribution is highly concentrated in the range of 0.88~0.94 kg·m-3, accounting for 79.57% of the occurrences. A low-density power gain coefficient is defined, and the critical wind speed at which the gain transitions from positive to negative is identified as 3.39 m·s-1, corresponding to a chord?based Reynolds number of approximately 4.49×105~4.69×105, which falls within the boundary-layer transition regime. A decreasing sigmoid model is established to describe the gain evolution, with a maximum fitting deviation of 3.77%. The influence of air density is wind-speed selective: in the start-up regime, lower density favors higher power and rotor speed; in the high?wind regime (from 9.5 m·s-1 up to the rated wind speed), higher density yields superior power output; and beyond the rated wind speed, moderate density results in the optimal conversion efficiency. The threshold for triggering the constant?power/constant?speed regimes varies with air density—higher density delays the onset of power saturation, whereas lower density delays rotor?speed saturation, thereby corroborating the control logic of “speed regulation first, power regulation thereafter.” The proposed model can be embedded into maximum power point tracking (MPPT) for density-adaptive torque compensation, with an estimated annual energy yield increase of approximately 1.69% in the low?wind regime. This study reveals the density?modulation mechanisms over the plateau and provides quantitative references for the optimized design of wind turbines at high-altitude sites.